From your description I wasn't entirely sure how you apply the hold-down forces, but here are a few thoughts:
Making the landing gear stiffer can't help. While it means that the vessel sinks less deep into the ground, the spring energy stored in the compressed landing gear is exactly the same as for the more compressible gear, and it will still all be released at once when the bolts are blown. The only reason a more compressible gear might behave better is because its eigenfrequency is lower, resulting in a more stable numerical solution.
The problem with applying a fixed hold-down force is that at equilibrium the landing gear acts like a compressed spring, catapulting the vessel up when the clamps are released.
The correct way to implement the hold down force imo would be to balance it with gravitational force and engine thrust:
1. while the rocket sits on the ground idly, no hold down forces are applied.
2. when the engines are firing up (assuming they are not at full thrust instantly) no hold down forces are applied until the thrust equals the weight of the launch stack. At this point, the landing gear will no longer be compressed, and does not store any more spring energy.
3. when engine thrust exceeds weight, you apply hold-down forces that exactly balance that difference, keeping the vessel on the ground with uncompressed gear
4. upon release of the clamps, the hold-down forces are removed instantly. The upward acceleration
should then be smooth, because the gear won't release any energy.
One potential problem with this approach is that during phase 3, there is no lateral surface friction, which means that the vessel might start sliding. To avoid that, you should probably increase the hold-down forces slightly, and define very high lateral friction coefficients for the relevant touchdown points. In that case, you probably also want to remove the hold-down forces more gradually, to avoid any residual spring effect.
In practice, I guess phase 3 won't last long (only until the engines have stabilised), so maybe the sliding problem won't be severe.
Of course, if your launch platform is defined as a vessel, then Dave's suggestion will be the easiest to implement, and it makes perfect physical sense. If you make the rocket thrust high enough, you would even be able to lift the launch pad! :lol:
Edit: Maybe the two-part force profile of steps 2 and 3 could be replaced by a single linear profile: hold-down force is zero at engine thrust zero, and the difference of engine-thrust minus weight at max thrust, and a linear function of engine thrust in between. This way, the landing gear will be gradually relaxed, and will be fully extended at clamp release. Plus, you would retain lateral friction all the way to release.
martins, I'd like your input on something. For Apollo, when the Saturn 1st stage engines ignite, there is a device to hold the vehicle down until thrust has built up to full. We simulated this in 2010-P1 by applying a large negative force to counteract the thrust. In the new versions the default "landing gear" is compressible so the hold-down force pushes the vehicle into the ground. I tried to define a non-compressible landing gear using the new API but that resulted in the vehicle being violently ejected into space instantly on first movement. Reducing the hold-down force results in early liftoff, and trying to balance it exactly results in the vehicle sinking into the ground during run-up. There are suggestions I should scrap the hold-down feature entirely and "start" the engines at T0, which would throw off the fuel usage numbers (and consequently 1st stage performance) and require a bunch of hacking to bring those back into the neighborhood of reality. Can you give us some kind of fix for this (a means of disabling the compression, or some other means of holding the stack down) or should I consider hold-down to be a permanently broken idea and redesign for an instant start at T0?